Canon 7D Noise?

Ok guys, I set my 7D to increments of 1:1 on the ISO. which as you know by this thread are 100,200,400,800,1600,3200 etc. I've had much better results using these ISO's BUT!!!!! even with 1:1 switched on if i choose auto ISO it doesn't stick to these ISO's and goes back to the 1:3 setting GGRRRR!!!

If I want the 7D to work in 1:1 increments do I have to set the ISO Manually?

Thanks guys.

Sent from my iPhone using TP Forums
 
Cpinch73 said:
Ok guys, I set my 7D to increments of 1:1 on the ISO. which as you know by this thread are 100,200,400,800,1600,3200 etc. I've had much better results using these ISO's BUT!!!!! even with 1:1 switched on if i choose auto ISO it doesn't stick to these ISO's and goes back to the 1:3 setting GGRRRR!!!

If I want the 7D to work in 1:1 increments do I have to set the ISO Manually?

Thanks guys.

Sent from my iPhone using TP Forums
Click to expand...

I don't know if the 7d chooses better than the 40d, but ISO is one of the parameters I definately want to be in control of.
 
Cpinch73 said:
Ok guys, I set my 7D to increments of 1:1 on the ISO. which as you know by this thread are 100,200,400,800,1600,3200 etc. I've had much better results using these ISO's BUT!!!!! even with 1:1 switched on if i choose auto ISO it doesn't stick to these ISO's and goes back to the 1:3 setting GGRRRR!!!

If I want the 7D to work in 1:1 increments do I have to set the ISO Manually?
Click to expand...

Yes. What's worse is that, in my limited experience of the feature, it seems predisposed to keep picking the worst offenders in the noise stakes. I usually shoot with full manual exposure, picking my own ISO and so on, but even when I do use Av or Tv mode I certainly don't let the camera pick the ISO. I'd be happy to use it if it worked as I would like, but it doesn't.
 
Cpinch73 said:
Ok guys, I set my 7D to increments of 1:1 on the ISO. which as you know by this thread are 100,200,400,800,1600,3200 etc. I've had much better results using these ISO's BUT!!!!! even with 1:1 switched on if i choose auto ISO it doesn't stick to these ISO's and goes back to the 1:3 setting GGRRRR!!!

If I want the 7D to work in 1:1 increments do I have to set the ISO Manually?

Thanks guys.

Sent from my iPhone using TP Forums
Click to expand...

Does the auto default to the setting for the iso spacing selected? thats normally 1/3 by default, try changing that and see if it makes any difference.
 
The ISO setting on my 7D is full stops, however if using Auto ISO it uses 1/3 stops, which is logical if you think about it. Using Manual you may set the shutter and aperture, so how is the camera going to get correct exposure using Auto ISO as the third part of the equation unless it uses 1/3 stops.
 
tdodd said:
Mine are set to 1:1 (if that's the right way to express it:thinking:) and I always select ISO manually. But like I say, that what works for me. Others may prefer to do it the other way.
Click to expand...

I decided to try out your Settings (1:1) ISO instead of (1:3) and noticed straight away the Improvement.

I Shoot mainly in Daylight with a 70-300 L. Lens.

Can also see the other points you make about if the Image is Jpeg etc.

Will also try out no ALO as mine is set to Standard.

Thanks for the Information.

Regards.
 
Last edited:
Thanks for an interesting thread with some very informative replies. I already use the camera setting Tim has recommended, but I can better understand why they work now. I still get some noise on the 7D, it’s not dreadful but more than I would hope for and seems more than in the pictures Tim posted on this thread.
Can I ask about exposing to the right and histograms please. For nature photography, I use between 400 and 1600 ISO generally, lower if the light allows, and I use Av normally.
Without applying any Exposure compensation, my histogram usually shows a graph starting at the left edge and finishing close to or on the right hand edge. The higher peaks tend to be biased towards the left though. I use this as an indication that I haven’t clipped either side.
If I apply EC to expose further to the right, because the graph was already on the right hand edge, the EC pushes the 255 end into overexposure. I think I understand the reasoning behind this method but am struggling to apply it because of the above reason. Am I looking at this wrong or have I misunderstood the principles.

With regard to the histogram, do you recommend brightness or RGB. I read somewhere that RGB is more accurate as it shows the levels on each channel, whereas with brightness it averages out all three channels. This results in the possibility that one channel could be clipped and another far from clipped, the mean average shown in the brightness histogram shows the levels to be unclipped even if one channel is clipped. I hope I have explained myself clearly here. I would appreciate your thoughts on this.
 
Trev, this article explains the reasons for ETTR - http://www.luminous-landscape.com/tutorials/expose-right.shtml

It is fine for the histogram to just touch the right hand side, that's perfect. It's also fine to have a peak very close to the right hand side. What you don't want is for the histogram to be piled up against the right hand side. Here are three examples I posted in another thread today, where I have achieved an ETTR exposure with just a hint of clipping, clipping that (a) is not a problem because in raw I can easily recover it; (b) because the areas concerned are insignificant and I don't really care if those tiny patches in the image do not contain texture or tonal detail.

image000sq.jpg


image000hr.jpg


image001mk.jpg


I achieve my ETR exposure with precision because I shoot with a manual exposure that is set by spot metering off the brightest parts of the scene at +3. +3 is the point where clipping just starts to appear. I use this technique with scenes of high dynamic range, snow scenes and when shooting birds or aircraft against a bright sky - I identify the brightest part of the sky through which I expect my subject to fly and spot meter it at +3, set my manual exposure and sit back and wait. I find this far easier than guessing at how much EC to dial in and twiddling the dial as the brightness of the subject and scene change. Here's one of my birdy examples....

image000uj.jpg


It needs a bit of tweaking to make it look nice, but I have maximised my exposure without losing any highlight detail at all.

Here's a plane....

image001cgi.jpg


Here I've got no clipping at all, but look how close my histogram is to the right hand edge. Another 1/3 stop would definitely give me some clipping, but note how dark and shaded the plane is. I wanted the high exposure in order to capture as much detail and as little noise in the shaded side of the aircraft as possible. Again some tweaking will be required to make it "look" nice, but as a raw image capture I have as much data as I could hope for.

As for in camera histograms, I use the RGB one, which is definitely the best one to use if you shoot to JPEG. For raw it is less important , but I prefer to have it there anyway.
 
Very useful information. I had a funny feeling there was something odd about the intermediate isos as I noticed exposure compensation values appearing that I knew I hadn't set.

Are higher isos pushed in the same way so some of them aren't quite 'real'?
 
Tim, thank you for such an excellent reply, very informative indeed, its appreciated :thumbs:
I wouldnt have dreamed of ETTR to the degree you have shown, I can see now how much to try for and understand your reasoning.
Can I just be sure I understood you with your settings please. Under the conditions you mentioned, you set EC to +3 first, and then meter the scene in the normal way and select the appropriate settings. You prefer manual metering, but I would assume this method would work equally well in Av or Tv.

Thank you again for taking the time to give such a detailed reply.
Trev
 
Suz, I don't know about Sony, but on some Canon cameras the very high ISOs, usually those marked as H or H1, H2 and also L (the 50 ISO setting on some Canons) are achieved by digital pushing and pulling. They are not real ISOs. I'm out of touch with whereabouts each camera switches from real to fake ISOs. I know the 30D and 40D topped out at 1600 ISO and 3200 ISO was faked, but I stopped using those bodies a couple of years ago. I think on everything I use today I am safe up to at least 3200 ISO. Beyond that I'm not sure, but I have only once shot at 6400 ISO in real world shooting so I'm happy with that.

The downside of using fake high ISOs is that the faked values might result in clipping, but will not result in additional data or lower noise. In effect they lose you a stop (or two) of highlight headroom. In theory you'd be better off setting the camera to the highest real ISO and underexposing by a stop (or two). Then you can use your editing software to fine tune the push rather than leave it to the camera to blindly double (or quadruple) your raw data values.
 
Trev4 said:
Tim, thank you for such an excellent reply, very informative indeed, its appreciated :thumbs:
I wouldnt have dreamed of ETTR to the degree you have shown, I can see now how much to try for and understand your reasoning.
Can I just be sure I understood you with your settings please. Under the conditions you mentioned, you set EC to +3 first, and then meter the scene in the normal way and select the appropriate settings. You prefer manual metering, but I would assume this method would work equally well in Av or Tv.

Thank you again for taking the time to give such a detailed reply.
Trev
Click to expand...

I leave the exposure mode set to Manual and I twiddle my exposure controls until the meter shows +3 when spot metering from the brightest part of the scene. I do not use autoexposure for this technique and therefore I do not use EC either. Also, for the precise control I seek I stick with manual exposure so the settings don't drift as the subject or scene changes slightly. I never understand why people would first meter in an autoexposure mode, and then transfer the exposure values into manual settings. Why not just set the manual exposure in the first place? Setting EC at +3 is no different to setting a manual exposure with the meter needle at +3. The end result is the same, but setting the exposure directly in manual mode is faster and less prone to error. Another attraction of setting the exposure manually is that many cameras don't offer EC values above +2, so how can you get to +3 in an autoexposure mode? You can't. With manual exposure on a +/-2 camera you first meter at +2, manually, and then simply increase the exposure by 1 stop above that.

Think of this example - you have a Lancaster bomber approaching. At first it is quite small in the viewfinder and doesn't have much impact on the exposure readings. As it gets closer it becomes larger and larger until it pretty much fills the frame. With autoexposure the exposure would keep changing as the plane grew and grew within the frame. I don't want that to happen. The exposure should remain constant not matter how small or large the plane. Once I set the exposure I want it to stay there. That's why I shoot manual exposure most of the time. I can change my composition and my exposure setting remains unaltered. That makes my life easier and my results more predictable.

It's the same thing as I track a bird in flight. The bird may fly past different backgrounds, maybe all sky, maybe all woodland, maybe some water, maybe a bit of each. Once I have my exposure set I do not want it to keep changing as I pan across the scene. So long as the lighting on the bird and the scene as a whole remains constant then that's how I'd like my exposure - constant.

Note, there is a bit more to doing doing all this. Often the approach I have described to ETTR exposure is as simple as I have described, but different scenes do sometime require different techniques. This might include scenes where there really are no highlights, such as a black dog running across a field of grass, with no sky within the frame. Then I think there is no need to chase the histogram all the way to the right. You just need to avoid underexposing the image. The real goal of all this is to never need to brighten a photo. Making it darker is OK. Having it leave the camera just perfect is OK too. Having it underexposed is (almost) never a good idea.

Here's an example where I have shot a little bright in order to capture full detail in the dog. The histogram does not touch the right hand edge and I don't care. My capture is bright enough that I don't need to brighten it further. This is exposed to the right, but not all the way to the right, if that makes sense....

image000cq.jpg



Another example where highlight metering does not work is where you do have important highlights, but the subject is (or highlights are) too small to meter from reliably, even with spot metering. Here again one needs to resort to different tactics to set the exposure. Here's one such image....

image000ni.jpg


Here the brightest parts of the swan are too small to meter. However, seeing the type of light hitting the swan I ignored the rest of the scene and set a manual exposure equivalent to a "Sunny 16" exposure. I didn't know for sure that I'd get the exposure spot on but reference to the histogram confirmed that I nailed it in one. If I'd needed to make a small adjustment that would be no problem. I can only imagine that trying to shoot this scene with autoexposure would be a right old faff and a case of repeated trial and error, with figures changing every time the composition changed just a little.
 
Fascinating read Tim, thanks so much for that.

I've only ever really used manual exposure in "difficult" situations such as shooting oncoming rally cars where the occasional dazzle from the lights plays havoc with auto-exposure but I think I'll make more of an effort to try it more often from now on and see how I do :)
 
Once again Tim, thank you. I'm pleased I did double check with you because I had got the wrong end of the stick. Its clear now and I'm certainly going to put that into practice. It make a lot more sense for BIF too, something I have struggled to get just right.
Cheers Tim, an excellent guide indeed :thumbs::clap:
 
You must log in or register to reply here.
Back
Top